Files
Moltbot/src/shared/text-chunking.ts
2026-02-19 07:01:54 +00:00

35 lines
1.0 KiB
TypeScript

export function chunkTextByBreakResolver(
text: string,
limit: number,
resolveBreakIndex: (window: string) => number,
): string[] {
if (!text) {
return [];
}
if (limit <= 0 || text.length <= limit) {
return [text];
}
const chunks: string[] = [];
let remaining = text;
while (remaining.length > limit) {
const window = remaining.slice(0, limit);
const candidateBreak = resolveBreakIndex(window);
const breakIdx =
Number.isFinite(candidateBreak) && candidateBreak > 0 && candidateBreak <= limit
? candidateBreak
: limit;
const rawChunk = remaining.slice(0, breakIdx);
const chunk = rawChunk.trimEnd();
if (chunk.length > 0) {
chunks.push(chunk);
}
const brokeOnSeparator = breakIdx < remaining.length && /\s/.test(remaining[breakIdx]);
const nextStart = Math.min(remaining.length, breakIdx + (brokeOnSeparator ? 1 : 0));
remaining = remaining.slice(nextStart).trimStart();
}
if (remaining.length) {
chunks.push(remaining);
}
return chunks;
}